Last month a homeowner in Sacramento sent a photo of her breaker box to a contractor and asked what it would take to add a Level 2 EV charger. The answer came back in three days: $4,800 for a full 200-amp panel upgrade, plus a utility service change that could take six weeks. She uploaded the same photo to Qmerit's Panel Insights, an AI trained on 269,000 residential electrical installations. It analyzed her panel in under a minute. Available circuit spaces, breaker types, tandem configurations, total capacity. The verdict: her existing 100-amp panel had 37 amps of headroom after accounting for every connected load. Plenty for a 40-amp EV charger circuit. No upgrade, no utility crew, no six-week wait.
She is not unusual. Seventy percent of homeowners already have enough panel capacity for an EV charger, according to ChargeRight founder Jason Walls, an IBEW-licensed master electrician who has made debunking unnecessary panel upgrades his entire business model.
What the AI Actually Sees
Panel Insights is not a chatbot guessing from your description. It is a computer vision model built by Qmerit in collaboration with Schneider Electric's AI Hub, a team of over 400 data scientists, using Microsoft Cognitive Services for the image recognition backbone. You open your breaker panel, aim your phone, and snap a single photo. The system identifies the panel manufacturer, counts available circuit spaces, spots tandem breakers that indicate a panel already running near its slot limit, reads the main breaker amperage, and catalogs the connected loads visible from breaker labels.
Then LoadCRE, which stands for Load Capacity Recommendation Engine, runs the numbers that your electrician is supposed to run but often doesn't. It applies NEC Article 220 load calculations against the panel data that Panel Insights extracted and determines exactly how much capacity remains for new loads: an EV charger, a heat pump, an electric water heater, an induction range. The process that traditionally required a licensed electrician to physically inspect your panel, catalog every circuit, and hand-calculate your load on a worksheet completes in seconds from a phone camera.
Qmerit CTO Manoj Puthenveetil has described the 10-month development process with a revealing detail: most of the time went into labeling the training data correctly, not building the model itself. His team brought in working electricians to sit alongside data scientists and argue about what mattered in a breaker photo, a process that exposed how much institutional knowledge had never been formalized in any training manual or inspection guide. Every correction a field electrician makes to a Panel Insights assessment gets fed back into the model automatically, creating a self-reinforcing feedback loop that has been refining its accuracy continuously since the system first went live with real installations.
Why Electricians Default to "Upgrade"
There's an uncomfortable truth behind the reflex to recommend a panel upgrade. It pays well. A 200-amp panel swap runs $1,300 to $3,000 for hardware and labor, according to This Old House and Angi. But that number is misleading because it rarely captures the full scope of work. When the utility needs to upgrade your service drop, replace the meter, and run new feed wires, you are looking at $3,000 to $10,000. In some markets, considerably higher. Schneider Electric SVP Richard Korthauer has publicly stated that a 400-amp utility service upgrade can add "$17,000 in builder development costs per home."
Recommending an upgrade when you could solve the problem with a $200 load management device or a simple NEC 220.82 Optional Method load calculation is not malice; it is a habit baked into decades of training programs that emphasized conservative sizing over cost-effective alternatives and never taught the Optional Method as a legitimate path. The NEC Standard Method (Article 220.82 is the Optional Method, a less conservative alternative that accounts for demand diversity) has been in the code for decades, but many electricians were never trained on it. They default to the conservative approach: add up every nameplate rating, assume everything runs simultaneously, conclude the panel is overloaded. Nobody checks. Real homes do not work that way. Your dryer, oven, water heater, and air conditioner do not all draw peak load at the same moment, and the Optional Method accounts for that statistical reality to produce a load figure that is often 30 to 40 percent lower than the Standard Method.
The $1,500 Alternative That Didn't Exist Two Years Ago
Even for the 30 percent of homes that genuinely lack capacity, a full panel upgrade is no longer the only answer. SPAN released the Drive in 2025, a two-circuit add-on module that costs $1,500 to $2,200 for the hardware. It bolts onto your existing panel with no rip-and-replace. The Drive manages two high-power 240V circuits dynamically: when your EV charger is pulling 40 amps and your dryer kicks on, it automatically throttles the charger to keep total demand under your panel's capacity. When the dryer stops, charging ramps back up. It is UL 3141 certified for managed EV circuits, and it enables 48-amp EV charging on a 100-amp or 125-amp panel. That is the full juice of a Level 2 charger on a panel that every electrician in America would have flagged for replacement two years ago.
| Option | Hardware Cost | Installed Cost | Panel Replacement? |
|---|---|---|---|
| 200A panel upgrade | $250–$500 | $1,300–$3,000+ | Yes |
| 200A + utility service change | $250–$500 | $3,000–$10,000+ | Yes + utility work |
| SPAN Drive (2 circuits) | $1,500–$2,200 | $2,000–$3,500 | No |
| SPAN Smart Panel (full) | $3,500–$4,500 | $6,000–$10,000 | Yes (smart replacement) |
| Schneider QO Smart Power Manager | Modular pricing | Varies by circuit count | No (add-on to QO panels) |
Schneider Electric is playing a different game with the Square D QO Smart Panel and its Smart Power Manager feature. Instead of replacing the whole panel, Schneider sells control relays that bolt into your existing Square D QO panel, the most common residential panel brand in America. The relays let the system shed loads when demand spikes. Your EV charger pauses for three minutes while the air conditioner compressor cycles through its startup draw, then resumes. You never notice. Schneider showed the system at IBS 2025, and Eaton dropped $75 million into SPAN in early 2026 to compete. The smart panel market is no longer speculative. It is an arms race between two of the largest electrical equipment manufacturers on the planet, and every residential panel sold in 2027 will carry the fingerprints of this fight.
NEC 2026 Changed the Rules
The 2026 edition of the National Electrical Code contains a provision that barely got covered in mainstream press but fundamentally changes the economics of residential electrification. For the first time, the NEC explicitly provides a code-compliant path for Power Control Systems to manage overload through active, real-time load monitoring instead of requiring panels sized for the theoretical maximum of every connected device running simultaneously.
The new PCS provisions span Articles 220, 625, and 706. A PCS-listed device, like the SPAN Drive, can now legally keep your total load below your panel's rated capacity by shedding lower-priority circuits when demand exceeds a programmable threshold. Previous editions of the code forced a simple binary: either your panel was big enough for the sum of all nameplate ratings (with demand factors applied), or you upgraded. Software could not substitute for copper; now it can, and the code says so explicitly.
There is a significant catch: NEC 2026 also requires that permanently installed EV chargers be installed by a "qualified person," language that NFPA voting members locked in during June 2025 by rejecting a motion that would have preserved the right to DIY installations. The 2026 NEC defines "qualified person" in terms that most states will interpret to mean a licensed electrician. So while the code now allows cheaper panel solutions, it simultaneously requires professional installation of the equipment that benefits from those solutions. Whether that net effect saves or costs the homeowner money depends entirely on local labor rates and utility service change costs in their market.
The Math Nobody Runs
A homeowner with a 100-amp panel, central air conditioning, an electric dryer, an electric range, and an electric water heater has a nameplate load that looks terrifying on paper. Add it up. Standard Method says you are pulling 150+ amps. Upgrade immediately. Case closed. Except it shouldn't be.
Optional Method tells a very different story. NEC 220.82 starts with the first 10 kVA of general load at 100 percent, then applies a 40 percent demand factor to everything else. The central AC counts at the larger of heating or cooling, not both. The result is typically 60 to 80 amps of calculated demand for the same house that the Standard Method claimed needed 150. Your 100-amp panel has 20 to 40 amps of headroom that nobody told you about. A 40-amp EV charger circuit fits. It fits without a panel upgrade, without a utility service change, and without a six-week wait for the utility to schedule a crew.
Running this calculation manually requires pulling out the NEC handbook, looking up demand factors in Table 220.42, cross-referencing with Article 220.82's specific provisions for existing dwellings, and doing arithmetic that most homeowners will never attempt. Qmerit's LoadCRE does it from a photograph. ChargeRight does it for free as part of their assessment service. Both exist because the gap between what the code actually allows and what electricians habitually recommend is wide enough to build a business inside.
The Strongest Case Against
AI panel assessment has real limitations that no amount of computer vision solves. Panel Insights reads breaker labels and visible hardware. It cannot see the condition of the bus bars behind the panel cover, cannot detect aluminum wiring that may require special treatment, cannot identify Federal Pacific Stab-Lok or Zinsco panels that should be replaced regardless of capacity because their breakers have documented failure-to-trip rates far exceeding industry standards. It cannot assess the age or condition of the service entrance cable from the utility transformer to your meter. A licensed electrician physically inside your panel can catch problems that will kill a computer vision model every time. No algorithm fixes that.
Load management devices also introduce a dependency that traditional panels do not have. If your SPAN Drive's firmware crashes at 2 AM while your EV charges and your dryer runs, your breaker trips. If the Wi-Fi goes down and the device loses its connection to load data from your other circuits, it may default to conservative throttling that leaves your car uncharged by morning. These are not hypothetical scenarios but rather the unavoidable tradeoff of replacing copper capacity with software intelligence: the software has to work, continuously, for years, in an environment where electricians have historically never needed to troubleshoot a firmware update.
And the electricians have a point about one thing. If you are already opening the panel to add an EV charger circuit, and your panel is 25 or 30 years old with corroded bus bars and breakers that feel loose when you toggle them, the smartest investment may genuinely be the full upgrade regardless of whether the load calculation says you have headroom. Capacity is not the only reason to replace a panel. Age, condition, and the shift from a gas-appliance home to an all-electric one over the next decade all argue for upgrading while the walls are already open and the electrician is already billing you. That is a legitimate call.
What to Do Before You Write That Check
If you have been quoted more than $2,000 to add an EV charger because your electrician says you need a panel upgrade, do three things before you approve the work.
First, photograph your open breaker panel (with the cover removed, dead front visible, all breakers labeled) and submit it to Qmerit's Panel Insights through their app or website. The assessment is free for homeowners getting quotes through their installer network, and if the AI says you have headroom, you have concrete data to bring when seeking a second opinion.
Second, ask your electrician specifically whether they ran an NEC 220.82 Optional Method load calculation or a Standard Method calculation. If they used Standard Method, ask them to rerun it using Optional. The difference is often 30 to 40 percent, and the Optional Method is fully code-compliant for existing dwellings. Any electrician who refuses to run the Optional Method calculation is either unfamiliar with it or uninterested in giving you the cheaper answer.
Third, get a quote for a SPAN Drive or equivalent load management device as an alternative to the panel upgrade. Hardware runs $1,500 to $2,200, installed cost $2,000 to $3,500 total. Compare that to the full panel upgrade quote. If the panel upgrade is $1,300 to $2,000 and your panel is aging, the upgrade may actually be the better value. If the upgrade quote is $4,000 or more because it requires utility service changes, the load management device saves you real money and gets your charger running weeks sooner.
Limitations of This Analysis
Qmerit's 70 percent figure comes from ChargeRight's analysis of their customer base and has not been independently verified by a peer-reviewed study. The actual percentage of homes that can accommodate an EV charger without a panel upgrade depends heavily on geographic region, housing stock age, existing electrical loads, and local utility service standards. Qmerit's Panel Insights accuracy claims have not been independently audited; the system's self-reported improvement loop is promising but unverified by third parties. SPAN Drive's UL 3141 certification covers the managed EV circuit specifically, but long-term field reliability data for residential load management devices deployed at scale does not yet exist in published form. NEC 2026's PCS provisions are adopted state by state, and as of June 2026, many jurisdictions have not yet adopted the 2026 edition. Cost figures cited reflect national averages and may vary significantly by market, particularly in high-cost-of-living areas where labor premiums can double the installed price of any option on the table.